The second cohort of Accelerator interns has completed their internship and these are some insights we got from their time with us.

1 Intern numbers

The cohort kicked off with 126 interns, with the majority being part of the Data science course. The completion rate was 88%. The interns who did not make it to the end either withdrew, or they were terminated by us.

Status Description
Active Intern was actively participating in the project
Excluded Intern stopped participating during the coursework phase and was not given the opportunity to do the internship
Withdrawn Intern decided to withdraw from the internship
Terminated We terminated the intern’s contract due to non-participation

Of the 126 interns we started with, 37 had incidents reported for either being MIA, needing leave or other reasons.

For a handful of the interns we were able to resolve their problems and integrate them into their teams again. In the event that we could not reach a solution we moved to terminate an intern’s contract. The most common reason for termination was non-participation.

2 Things we did differently

2.1 Workshops

We started the internship with a 2 week ‘orientation’ where we focused on project allocations and workshops. The main aim for the workshops was to fill gaps we recognised from the previous internship and highlight a few areas of importance that the interns needed to pay careful attention to. These are the skills we focused on for this cohort:

Workshop Purpose
Systems thinking Thinking about how work works & problem solving
Git fundamentals Emphasise best practices and how to collaborate with others properly
Writing How to communicate your technical work in writing & collaborate with others on a written piece of work.
Professionalism Start out your new job on the right foot and build a good reputation for yourself

The attendance for the workshop was lower than expected, on a good day we had 97 attendees on a call. A possible explanation for this is that a lot of our interns had jobs and were unavailable during the day.

2.2 Intern commitments

We collected information about what commitments the interns have before we started with project work. The aim of this was to have a split between part-time and full-time interns but because of the number of interns with commitments it did not end up being possible. It was valuable to share with mentors so that they can consider availabilities for meetings during the week.

2.3 Project preferences

Interns got the opportunity to choose their own projects. This was exciting and gave them the opportunity to work on things they were passionate about.They did not get the opportunity to do a lot of research on the topics but we provided enough information for them to have a general idea of what each project is about. We started off with 17 projects and the final number that got allocated is 13. This depended mostly on intern preference. The projects that had less than 4 people will be considered for the next cohort.

2.3.1 Team allocations

Teams were allocated based on project preference, commitments, Academy marks and the required number of data engineers/data scientists for each project. We started off with 20 teams working on 13 projects, and by the end we had 18 teams as 2 got merged due to non-participation. The teams were created with drop off in mind. Each team had between 6 and 7 members so that in the event that someone drops out, the team does not struggle because too few people are left.

2.4 Managing non-attendance

The biggest problem that mentors struggled with in the previous cohort was interns not participating and not having clear directions on how to handle this. For this cohort we had a policy for non-attendance that details what mentors should do if an intern doesn’t attend meetings and what timelines to follow. we also had an incident report database as a centralised place for Zintle to have visibility on who needs to be followed up on. This worked to a certain extent but it needs a bit of refinement for the next cohort. The policy does not take into account the effort that goes into re-integrating a team member after they’ve been gone for a certain period of time. Things move very quickly with the project and the policy needs to take this into account.

2.5 Project management

All the project management was centralised on Notion and each team started with a standard template. The teams struggled at first and this was a combination of using a new tool (Notion) and there being gaps in their knowledge. To solve for the first problem we created a how-to video to help with Notion basics. To help with the second problem, we provided feedback to the teams with specifics on where they can improve. These are some of the gaps we identified:

  • Not recording meeting minutes properly
  • Not using correct methods for project management like task management, keeping track of everything in one place and assigning the right people to tasks.

Being able to identify the gaps was valuable for providing feedback to the Academy. The current cohort of students has the opportunity to improve on this early on.

2.6 Mentor roles

Mentor role Description
Specialists Commit to 1 hour a week and support the team with a specific skill
Once-off Offer a masterclass or a Workshop
Team mentors Dedicated team mentor providing technical & professional guidance

2.7 Centralised git repository

We created repositories for all the teams and centralised their management. We then took screenshots of each repo at the end of each sprint to get a sense of how the teams were collaborating.

2.8 AWS

With the help of Muzi, we created an R&D account and the interns got new credentials to access the services. To start off these are the services they were given access to:

Service Resource type
EC2 t2 or t3
S3 S3 Bucket
RDS db.t3.micro
DynamoDB table

To get access to additional services, teams had to motivate for this and get approval. May was the most expensive month and EC2 instance accounted for the majority of the costs. The interns were closer to the end of their projects at this point and were busy deploying and testing their apps in the cloud.

3 Lessons learned

  • Explorers can participate in the internship in more than 1 way and making this possible resulted in a lot of volunteers.
  • Being more flexible with meeting times to take jobs and families into considerations can lead to better outcomes as the interns are able to let us know when they are available to participate.
  • Consistent thrive support for both mentors and interns is very valuable as it helps them realise they’re not alone.
  • Letting interns choose their own projects is a great way to start the internship since it places the decision in their hands.

4 Plans for the next cohort